Linear and Parallel Learning of Markov Random Fields
Machine Learning
2014-02-06 v4 Machine Learning
Abstract
We introduce a new embarrassingly parallel parameter learning algorithm for Markov random fields with untied parameters which is efficient for a large class of practical models. Our algorithm parallelizes naturally over cliques and, for graphs of bounded degree, its complexity is linear in the number of cliques. Unlike its competitors, our algorithm is fully parallel and for log-linear models it is also data efficient, requiring only the local sufficient statistics of the data to estimate parameters.
Cite
@article{arxiv.1308.6342,
title = {Linear and Parallel Learning of Markov Random Fields},
author = {Yariv Dror Mizrahi and Misha Denil and Nando de Freitas},
journal= {arXiv preprint arXiv:1308.6342},
year = {2014}
}